In the previous article, I introduced the basic aging curve for men's and women's tennis players to give a general idea of how performance varies with age. Using a one-size-fits-all curve for adjustments to all players is better than nothing, but it's not very useful for predicting a specific player's change in performance, nor does it give any insight into how and why tennis players decline as they get older. To do that, aging curves need to be developed for similar groups of players and understanding how and where each set of curves are different. Very basic groupings of players and their aging curves are shown below.
Grouping players is a huge task in itself. There's no shortage of potential statistics for classifying types of players (second-serve percentage, ace rate, average rally length, etc.), but adjusting for quality of competition and surfaces alone complicates things really fast. Instead, I opted to start with the simplest approach to grouping. I asked a single question: What is their best surface? I classified players as being hard-court players, clay-court players, or surface-netural players using the Advanced Baseline surface factors. Using the previously described method, I constructed aging curves for each type of player. The men's and women's aging curves are shown below for each type of player.
It's not super-clear from this graph, but hard-court players have a little bit of a softer landing than clay-courters after 30 to 32. My guess is that it's probably because the advantages hard-courters have (mainly their big serves) are less prone to breakdown over time. John Isner and Ivo Karlovic aren't going to stop being tall as they get older, so their serves will still retain their potency over time. Clay-courters, on the other hand, derive their edge more from cross-court movement and topspin, which put a lot more stress on the joints and requires more endurance. Neither of those two traits age all that well, so it makes sense clay court players have much more of a cliff after 32.
Women's hard-court players are the most perplexing subgroup I've analyzed so far. Their aging curve is just flat for most of the time, until the dropoff starts around age 27. I'm not convinced I haven't missed anything yet, but if this holds up to be true, it could mean that women's hard-courters derive their edge more from natural ability than development as a player. But I'm not sure why that's not the case for men's hard-courters. Surfaces definitely play out differently in the men's and women's games; understanding exactly how is still in its infancy from a stats perspective.
(By the way, that uptick at age 32 that skews the hard-court graph? That's Serena Williams breaking the aging curve by herself. Granted, there isn't a lot of data for women's hard-court players over 30 so small sample sizes and all that, but it's still a Rodman-esque level of outlier.)
In the next two posts, I'll apply these specific aging curves to the top men's and women's tennis players to see if they can help clarify what the top of the tours will look like in a couple years.